1. Field of the Invention
The present invention relates to a file search method of searching for a file with high relevance to a file which a user desires to obtain, and a system therefore. In particular, the invention relates to a file search method of searching for a file which a user desires to obtain among an enormous number of files stored in a user terminal device or a file server, and a system therefore.
2. Description of the Related Art
In consequence of the spread of the network environment including information devices and the Internet, a user has a large number of electronic files such as documents and images stored in a user terminal device or a file server. When the user tries to find out a desired file from the great number of files, a keyword is input to perform a search. In such a search, all the files including the keyword are displayed as the search results. Thus, the user searches for the desired file among the many search results.
Up to now, various techniques have been proposed for instantly finding out a desired file when the user performs a search.
For example, Japanese Unexamined Patent Application Publication No. 2001-147936 discloses a system for assigning a relevant value indicating the degree of relevance between documents in terms of citing and cited relation for the respective documents to search for a document on the basis of the assigned relevant value.
In addition, for example, Japanese Unexamined Patent Application Publication No. 2004-206631 discloses a system for registering a value in a database (DB) as the importance and performing a search on the basis of the importance. The value is obtained by setting information indicating the number of appearing times of a certain word in a document to be reflected by the characteristic amount as a weight based on the appearing status of the certain word in a document group.
According to the above-mentioned conventional search systems, for document files on which a cited document such as a patent document is described, a citation relation between document files is extracted by analyzing the description of the cited document to calculate the importance of each document file on the basis of the citation relation. When a file, in which the conventional search system is locally employed, is used as a search target, a description for performing the association in the document file is required. However, no citation description generally exists in the file in which the search system is locally employed, and thus such search systems cannot be applied to many files. Furthermore, as the citation relation depends on the description method, it is difficult to automatically extract the citation description.